Hex me

A few weeks ago I made the 'ratchet hubs' that went between the wheels/prop and the transmission. The hubs themselves fit nicely on our aluminum axles (as they were designed to do), but I had to come up with some way to connect the half shafts into the ratchet hub. Tums out that 3/4" hex bar slides nicely into the axles.

I drilled out the hex for weight savings. Also on the prop shaft we may eventually run a control rod through the hollow shaft for variable pitch control. If I didn't also drill the hex out I would be unable to run this control rod through the assembly.


As previously mentioned, my hex bar was 3/4", but my ratchet takes an 11/16" hex head so again a bit of milling to make them match up.

All that is left to do is secure the hex stub inside the shaft on the right. Perhaps a roll pin or two. I'll work on that tomorrow evening.

Odds and ends

A bit of pickup work this evening -- odds and ends that will allow the final assembly of the prop hub and the installation of the prop spars:

I milled off the tops of the stop and load cell contact bolts. The hub can now freely move fore and aft on the roller even under load. Also adjusted all the stops.



I installed the three load cell over-load limit set screws in the thrust plate. These insure we don't deform the load cells

I drilled, tapped and installed the screws in the prop spar 'dog bones' that hold the bearings in place. These bearings are inset into the bones and generally will be pulled into their sockets -- the screws only hold the bearings in place when not under load.

I also sliced the small bronze sleeves that will act as spacers -- holding the spar collars against the bearings but away from the bones. This will allow the spar to rotate freely for pitch adjustment.

Oh for a CNC mill

The last major part to the prop hub was the torque paddle. I cranked that out last night.

On each end of the paddle is a roller assembly that will allow the prop hub to compress the thrust load cells without binding up on the torque arms.

The assembled paddle.

The paddle installed.

Getting closer.

A shot of the thrust plate with the three load cells mounted radially. There is a thrust bearing which will press against these cells.


The thrust and torque cells installed. There is major piece missing on the rear of the assembly -- the 'paddle' which transfers torque from the axle to the hub through those two load cells. That component is my next milling project ... this evening.

Where it all comes together

First assembly of most of the basic hub components including the thrust plate shown in the last picture. The load cell brackets and stops can be seen on the rear place (which is facing outward in the first two pics).

Throw me a freakin' bone here

Drilling and tapping the central spar 'dog bones'

He aint't heavy ...

Milling out the load cell brackets and modifiying the load cell itself.

Circle this

Near twin prop hub plates fresh off the milling machine.

Talk to me

The instrumentation is starting to take shape. Given all the work that’s going into this project, we want to be absolutely certain to collect quality data. This is somewhat complicated by the fact that there are effectively three different data collection centers – the ground, the cart chassis, and the propeller.

Because the propeller is spinning, we chose to use a radio link to send its data to a computer on the cart’s chassis to be logged in real-time. We found a very slick device made by Texas Instruments. The EZ430-RF2500 includes two integrated devices that include a microprocessor, radio, and data acquisition capabilities. Unfortunately, no one on the team knows how to design low-level circuits or program such devices. However, there’s no better opportunity to learn.

With a fair amount of help from friends, we’ve muddled our way through to developing a prototype circuit to read and amplify the output of a load-cell, and get it into the 430. With more help from friends, we figured out how to program both ends of the 430 to read the signals from the load-cells, transmit them to the chassis, read them on the other end, and stream them into the laptop to be logged (and displayed to the pilot). Finally, we had to figure out how to read a serial stream in Windows since they no longer make laptops with serial ports and my old DOS serial routines can’t talk to a USB serial converter.

Having gotten this far, we’re feeling good about our ability to collect the data we need – not just for “the event”, but also for testing component performance (e.g. transmission, rolling resistance, propeller torque and thrust…). This will allow us to optimize the prop pitch, gear ratio, and other parameters before we attempt actual trials.

Fortunately, one of our friends that we met on the internet forums (Christian Klippel) has designed a really slick circuit and PC board to replace the prototype circuit we’ve cobbled up for proof of concept. We’ll be sending that out for fabrication in the next day or so. That should cover our instrumentation of the propeller. This will give us thrust, torque, and prop-pitch at all times.

Next comes the instrumentation on the chassis. Again we’ll use the 430 to collect load-cell data (for things like the force on the drive axle), but will need to develop a new bag of tricks for measuring things like prop RPM, wind speed, and wind direction. More on this as it develops.

Continental drift and propeller load testing

In addition to working on the instrumentation, in his garage, Rick is putting the finishing touches on the propeller. The blades are clamped by their spars and are fully cantilevered from these brackets.

Rick's garage is also more like 'toy central', with flexible flying craft of all sorts along with a collection of harnesses etc. -- too much stuff in too small of a space.

There is a shelf (too) full of stuff that resides over the area where the prop blades currently are being finished.



Here is the shelf after it decided to 'drift' rather dramatically.

Here is the stuff that was on the drifting shelf as it bounced off the prop blades.

Here is the very minor dent in the blade from the drifting shelf. Load test passed.

On second thought ...

After thinking about how to handle the torque transfer from the drive axle to the sprocket cassette, I decided to go the more difficult but more elegant route. I had intended to use a kart hub to handle the torque load(previous post) and afix the cassette housing to it -- also sleeving the shaft a bit since the cassette's ID was slightly bigger than our axle.

After sleeping on it I decided that I would build an indexing jig and spline the shaft itself. By custom milling both the keyway and the keystock, I could size the keys to handle both the radial loads and the torque.

I started with a round piece of plyood and using my previously contructed hub holder created a indexed wheel.

I set up a little marker system off to the side of the mill table and used this to locate each spline location on the axle. Milling ensued.

Custom milling the keystock.

I left enough room in the keyways to move the sprocket group back and forth for perfect alignment. The shaft collars are only used to keep the sprocket from sliding back and forth on the axle.

Next ...

Our first tests with the vehicle will be on a dynamometer where we collect data on the prop at static thrust (wind speed). We will also use the vehicle as a 'dynamic prop stand' by running it (push or pull?) up and down the runway in still air at different speeds, collecting data from our instrumented hub.

Before we build our two spools for the spool transmission, we want to make sure we get our gear ratios optimized. I ordered up a large recumbent bike sprocket for the prop shaft (65t) and purchased a 9 speed cassette for the drive axle. This will give us a selection of ratios to work with. Additionally, the chain drive will be our less efficient (we believe) backup transmission if our spool drive adventures don't play out.

The next challenge will be to mate the cassette on the left to the hub on the right. Not exactly sure yet how to get that done ... thinking cap on.

One way out

We don't have a differential on the drive axle -- too heavy for our tastes and besides, the vehicle is designed to pretty much stick to a straight line. We still need a way however to ensure we aren't binding as we drift around off the straight and narrow. We also need a one way device on the prop shaft to keep the prop from backdriving the transmission (especially the spool transmission) if we get a sudden wind shift.

The solution was three ratchet hubs that will mate up to our kart axles. Of course as has been the case with most of the parts on this vehicle, no one seems to stock DDWFTTW ratchet hubs anymore. After about 16 hours (12 in the first one and 2 hrs each in the others) I finally finished them up this evening

Construction details in previous post.

When you have more time than money

(Or to put it another way ... what to do when DDWFTTW Vehicles R Us is out of stock on ratchet hubs)

To make the ratchet hubs, we again started with standard kart hub blanks.


Previously, I searched around and found a ratchet box wrench that I felt would handle our torque needs and was also reversable. We'll use the reversing feature to allow us to tow the cart around free from the transmission.

After putting the hub on a diet, I milled out a shape that matched the shape of the wrench head.

I grabbed a square swatch of aluminum for a lid -- drilled, tapped and cap screwed it on and then milled it round to match the hub.

I ground off the handle of the box wrench and installed three set screws to use for centering the wrench in the hub.

Once I fit, centered and locked in the wrench head with the set screws, I mixed up a small batch of epoxy and heated up the assembly (so the epoxy would flow). I let the epoxy fill up the space between the hub and the wrench head to make a nice, tight, torque resistant fit.

Here's a shot of the three in a progressive state of assembly.